Human Develoment Report 2023/2024 technical notes

Calculating the human development indices—graphical presentation

Human Development DIMENSIONS Long and healthy life Knowledge A decent standard of living
Index (HDI)

INDICATORS Life expectancy at birth Expected years Mean years GNI per capita (PPP $)
of schooling of schooling

DIMENSION Life expectancy index Education index GNI index

INDEX

Human Development Index (HDI)

Inequality-adjusted DIMENSIONS Long and healthy life Knowledge A decent standard of living
Human Development
Index (IHDI)
INDICATORS Life expectancy at birth Expected years Mean years GNI per capita (PPP $)
of schooling of schooling

DIMENSION Life expectancy Years of schooling Income/consumption

INDEX

INEQUALITY- Inequality-adjusted Inequality-adjusted Inequality-adjusted

ADJUSTED life expectancy index education index income index
INDEX

Inequality-adjusted Human Development Index (IHDI)

Gender Development Female Male

Index (GDI)

DIMENSIONS Long and Standard Long and Standard

healthy life Knowledge of living healthy life Knowledge of living

INDICATORS Life expectancy Expected Mean GNI per capita Life expectancy Expected Mean GNI per capita
years of years of (PPP $) years of years of (PPP $)
schooling schooling schooling schooling

DIMENSION
INDEX Life expectancy index Education index GNI index Life expectancy index Education index GNI index

Human Development Index (female) Human Development Index (male)

Gender Development Index (GDI)

(continued)

TECH NICAL NOTES 1–6 1

Calculating the human development indices—graphical presentation (continued)

Gender Inequality DIMENSIONS Health Empowerment Labour market

Index (GII)
INDICATORS Maternal Adolescent Female and male Female and male shares Female and male
mortality birth population with at least of parliamentary seats labour force
ratio rate secondary education participation rates

DIMENSION Female reproductive Female empowerment Female labour Male empowerment Male labour
INDEX health index index market index index market index

Female Male
gender index gender index

Gender Inequality Index (GII)

Multidimensional DIMENSIONS Health Education Standard of living

Poverty Index (MPI)

INDICATORS Nutrition Child mortality Years School Cooking Sanitation Drinking Electricity Housing Assets
of schooling attendance fuel water

POVERTY Intensity Headcount

MEASURES of poverty ratio

Multidimensional Poverty Index (MPI)

Planetary pressures– Carbon dioxide emissions Material footprint

adjusted Human per capita (production) per capita
INDICATORS
Development Index
(PHDI)

Human Development Adjustment factor for

Index (HDI) planetary pressures

Planetary pressures–adjusted HDI (PHDI)

2 HUMA N DEVELOP MENT R EP ORT 2023/2024

Technical note 1. Human Development Index
The Human Development Index (HDI) is a summary The justification for placing the natural zero for life
measure of achievements in three key dimensions of expectancy at birth at 20 years is based on historical
human development: a long and healthy life, access to evidence that no country in the 20th century had a
knowledge and a decent standard of living. The HDI is life expectancy at birth of less than 20 years (Mad-
the geometric mean of normalized indices for each of dison 2010; Oeppen and Vaupel 2002; Riley 2005).
the three dimensions. This technical note describes the Maximum life expectancy at birth is set at 85, a realis-
data sources and the steps to calculating HDI values. tic aspirational target for many countries over the last
30 years. Due to constantly improving living condi-
Data sources tions and medical advances, life expectancy at birth
in several economies already exceeded or came very
• Life expectancy at birth: UNDESA (2022). close to 85 years in 2022: 86.9 years in Monaco and
• Expected years of schooling: CEDLAS and World 84.8 years in Japan.
Bank (2023), ICF Macro Demographic and Health Societies can subsist without formal education, jus-
Surveys (various years), UNESCO Institute for tifying the education minimum of 0 years. The maxi-
Statistics (2023) and United Nations Children’s mum for expected years of schooling, 18, is equivalent
Fund (UNICEF) Multiple Indicator Cluster Surveys to achieving a master’s degree in most countries. The
(various years). maximum for mean years of schooling, 15, is the pro-
• Mean years of schooling: Barro and Lee (2018), jected maximum of this indicator for 2025.
ICF Macro Demographic and Health Surveys (var- The low minimum value for gross national income
ious years), OECD (2023), UNESCO Institute for (GNI) per capita, $100, is justified by the considera-
Statistics (2023) and UNICEF Multiple Indicator ble amount of unmeasured subsistence and nonmar-
Cluster Surveys (various years). ket production in economies close to the minimum,
• GNI per capita: IMF (2023) and UNDESA (2023). which is not captured in the official data. The maxi-
• United Nations Statistics Division (2023) and mum is set at $75,000 per capita so that addition-
World Bank (2023). al income above the maximum does not directly
contribute to a country’s HDI value or improve its
Steps to calculate Human ranking. Only five countries (Liechtenstein, Qatar,
Development Index values Singapore, Ireland and Luxembourg) exceeded the
$75,000 income per capita ceiling in 2022.
There are two steps to calculating HDI values. Having defined the minimum and maximum val-
ues, the dimension indices are calculated as:
Step 1. Creating the dimension indices
actual value – minimum value . (1)
Dimension index =
maximum value – minimum value
Minimum and maximum values (goalposts) are set
in order to transform the indicators expressed in dif- For the education dimension, equation 1 is first
ferent units into indices between 0 and 1. These applied to each of the two indicators, and then the
goalposts act as “the natural zeros” and “aspiration- arithmetic mean of the two resulting indices is taken.
al targets,” respectively, from which component indi- Using the arithmetic mean of the two education in-
cators are standardized (see equation 1 below). They dices allows perfect substitutability between expect-
are set at the following values: ed years of schooling and mean years of schooling,
which seems to be right given that many developing
Dimension Indicator Minimum Maximum countries have low school attainment among adults
Health Life expectancy at birth (years) 20 85 but are very eager to achieve universal primary and
Education Expected years of schooling (years) 0 18 secondary school enrolment among school-age
Mean years of schooling (years) 0 15 children.
Standard of living GNI per capita (2017 PPP$) 100 75,000

TECH NICAL NOTES 1–6 3

 Because each dimension index is a proxy for capa- indicator (entirely or partly), the Human Develop-
bilities in the corresponding dimension, the transfor- ment Report Office calculates it by converting GNI
mation function from income to capabilities is likely per capita in local currency from current to constant
to be concave (Anand and Sen 2000)—that is, each terms using two steps. First, the value of GNI per cap-
additional dollar of income has a smaller effect on ex- ita in current terms is converted into PPP terms for
panding capabilities. Thus for income the natural log- the base year (2017). Second, a time series of GNI per
arithm of the actual, minimum and maximum values capita in 2017 PPP constant terms is constructed by
is used. applying the real growth rates to the GNI per capita
in PPP terms for the base year. The real growth rate
Step 2. Aggregating the dimensional indices is implied by the ratio of the nominal growth of GNI
per capita in current local currency terms to the GDP
The HDI is the geometric mean of the three dimen- deflator.
sional indices: For several countries without a value of GNI per
capita in constant 2017 PPP terms for 2022 reported
HDI = (IHealth . IEducation . IIncome ) 1/3
in the World Development Indicators database, real
growth rates of GDP per capita available in the World
Example: Nigeria (2022 HDI) Development Indicators database or in the Interna-
tional Monetary Fund’s Economic Outlook database
Indicator Value are applied to the most recent GNI values in constant
Life expectancy at birth (years) 53.6 PPP terms.
Expected years of schooling (years) 10.5 Official PPP conversion rates are produced by the In-
Mean years of schooling (years) 7.6 ternational Comparison Program, whose surveys peri-
Gross national income per capita (2017 PPP $) 4,755 odically collect thousands of prices of matched goods
and services in many countries. The last round of this
Note: Values are rounded.
exercise refers to 2017 and covered 176 economies.
53.6 – 20
Health index = = 0.517
85 – 20 Human development categories
10.5 – 0 The 2014 Human development Report introduced a
Expected years of schooling index = = 0.583
18 – 0
system of fixed cutoff points for the four categories of
human development achievements. The cutoff points
7.6 – 0
Mean years of schooling index = = 0.506 (COP) are the HDI values calculated using the quar-
15 – 0
tiles (q) from the distributions of the component indi-
0.583 + 0.506 cators averaged over 2004–2013:
Education index = = 0.545
2
COPq = HDI (LEq , EYSq , MYSq , GNIpcq), q = 1,2,3.
ln(4,755) – ln(100)
Income index = = 0.583
ln(75,000) – ln(100)
For example, LE1, LE2 and LE3 denote three quartiles
of the distribution of life expectancy at birth across
Human Development Index = countries.
(0.517 . 0.545 . 0.583)1/3 = 0.548 This Report keeps the same cutoff points on the
HDI for grouping countries that were introduced in
Methodology used to express income the 2014 Report:

The World Bank’s 202 3 World Development Indi- Very high human development 0.800 and above

cators database contains estimates of GNI per cap- High human development 0.700–0.799

ita in constant 2017 purchasing power parity (PPP) Medium human development 0.550–0.699

terms for many countries. For countries missing this Low human development Below 0.550

4 HUMA N DEVELOP MENT R EP ORT 2023/2024

 Human Development Index aggregates expectancy at birth and GNI per capita are weight-
ed by total population, expected years of schooling
Aggregate HDI values for country groups (by human is weighted by population ages 5–24 and mean years
development category, region and the like) are cal- of schooling is weighted by population ages 25 and
culated by applying the HDI formula to the weight- older.
ed group averages of component indicators. Life

Technical note 2. Inequality-adjusted Human Development Index

The Inequality-adjusted Human Development Index Inequality in the distribution of HDI dimensions is
(IHDI) adjusts the Human Development Index (HDI) estimated for:
for inequality in the distribution of each dimension • Life expectancy, using data from complete life ta-
across the population. It is based on a distribution- bles provided by UNDESA (2022). Mortality rates
sensitive class of composite indices proposed by Fos- and other information are available for each age (0,
ter, Lopez-Calva and Szekely (2005), which draws on 1, 2, 3, …. 100+). This is a major update: in previous
the Atkinson (1970) family of inequality measures. Reports the inequality in the distribution of life
It is computed as a geometric mean of inequality- expectancy was computed based on abridged life
adjusted dimensional indices. tables, which presented information on mortality
The IHDI accounts for inequalities in HDI dimen- patterns over age intervals (0–1, 1–5, 5–10, … 100+).
sions by “discounting” each dimension’s average • Mean years of schooling, using household sur-
value according to its level of inequality. The IHDI veys data harmonized in international databas-
value equals the HDI value when there is no inequal- es, including CEDLAS and World Bank (2023),
ity across people but falls below the HDI value as in- Eurostat’s European Union Statistics on Income
equality rises. In this sense the IHDI measures the and Living Conditions (2023), ICF Macro De­
level of human development when inequality is ac- mographic and Health Surveys (various years), LIS
counted for. (2023), United Nations Children’s Fund Multiple
Indicator Cluster Surveys (various years) and
Data sources UNESCO Institute for Statistics (2023).
• Income, using data from the United Nations
Since the HDI relies on country-level aggregates such University’s World Income Inequality Database
as national accounts for income, the IHDI must draw (UNU-WIDER 2023).
on additional sources of data to obtain insights into A full account of data sources used for estimating
the distribution. The distributions are observed over inequality for the 2022 IHDI is available at https://hdr
different units—life expectancy is distributed across a .undp.org/inequality-adjusted-human-development
hypothetical cohort, while years of schooling and in- -index#/indicies/IHDI.
come are distributed across individuals.
Steps to calculate Inequality-adjusted
Human Development Index values

There are three steps to calculating IHDI values.

TECH NICAL NOTES 1–6 5

 Step 1. Estimating inequality in the three Step 3. Combining the dimension indices
dimensions of the Human Development Index
The IHDI is the geometric mean of the three dimen-
The IHDI draws on the Atkinson (1970) family of in- sion indices adjusted for inequality:
equality measures and sets the aversion parameter
ε equal to 1.1 In this case the inequality measure is IHDI = (I Health
* . I* . I * )1/3 =
Education Income

A = 1– g/μ, where g is the geometric mean and μ is the [(1 – A .

) (1 – A
Health
.
) (1 – A
Education Income
)]1/3 . HDI.
arithmetic mean of the distribution. This can be writ-
ten as: The loss in HDI value due to inequality is:

X1 …Xn Loss = 1 – [(1 – AHealth) . (1 – AEducation) . (1 – AIncome)]1/3.

n

Ax = 1 – – (1)
X

Difference in Inequality-adjusted
where {X1, … , Xn} denotes the underlying distribution Human Development Index rank and
in the dimension of interest. Ax is obtained for each Human Development Index rank
variable (life expectancy, mean years of schooling
and disposable household income or consumption Difference in ranks on the IHDI and the HDI is calcu-
per capita). lated only for countries for which both an IHDI value
The geometric mean in equation 1 does not allow and an HDI value are calculated. Due to data limita-
zero values. For mean years of schooling one year tions, IHDI values are calculated for a smaller set of
is added to all valid observations to compute the countries. To calculate the IHDI rank difference from
inequality. the HDI rank, HDI ranks are recalculated for the set
of countries for which an IHDI value is calculated.
Step 2. Adjusting the dimension indices for inequality
Coefficient of human inequality
The inequality-adjusted dimension indices are ob-
tained from the HDI dimension indices, Ix, by multi- An unweighted average of inequalities in health, ed-
plying them by (1 – Ax), where Ax, defined by equation ucation and income is denoted as the coefficient of
1, is the corresponding Atkinson measure: human inequality. It averages these inequalities using
the arithmetic mean:
I *x = (1 – Ax ) . Ix . AHealth + AEducation + AIncome
Coefficient of human inequality = .
3
The inequality-adjusted income index, I , is *
income
based on the index of logged income values, I income*, When all inequalities in dimensions are of a similar
and inequality in income distribution computed magnitude, the coefficient of human inequality and
using income in levels. This enables the IHDI to ac- the loss in HDI value differ negligibly. When inequal-
count for the full effect of income inequality. ities differ in magnitude, the loss in HDI value tends
to be higher than the coefficient of human inequality.

Notes on methodology and caveats

The IHDI is based on the Atkinson index, which satis-

fies subgroup consistency. This property ensures that
improvements (deteriorations) in the distribution of
human development within only a certain group of
the society imply improvements (deteriorations) in
1 The inequality aversion parameter affects the degree to which lower achieve-
ments are emphasized and higher achievements are de-emphasized. the distribution across the entire society.

6 HUMA N DEVELOP MENT R EP ORT 2023/2024

 The main disadvantage is that the IHDI is not as- Example: Jordan (2022 IHDI)
sociation sensitive, so it does not capture overlapping
inequalities. To make the measure association sensi-
tive, all the data for each individual must be available Inequality Inequality-
Dimension measure adjusted index
from a single survey source, which is not currently Indicator Indicator index (A)a (I*)
possible for a large number of countries. Life expectancy (years) 74.2 0.8341 0.093 (1–0.093) . 0.8341
= 0.7566

Expected years of schooling (years) 12.6 0.7015 — —

Mean years of schooling (years) 10.4 0.6964 — —

Education index 0.6990 0.154 (1–0.154) . 0.6990

= 0.5914

Gross national income per capita 9,295 0.6846 0.241 (1–0.241) . 0.6846
(2017 PPP $) = 0.5196

Inequality-adjusted Human
Human Development Index Development Index

(0.8341 . 0.6990 . 0.6846)1/3 = 0.736 (0.7566 . 0.5914 . 0.5196)1/3 = 0.615

Loss due to inequality (%): Coefficient of human inequality (%)

100 . (0.093 + 0.154 + 0.241)

(
100 . 1 –
0.615
0.736)= 16.4
3
= 16.3

Note: Values are rounded.

a. Inequalities are estimated from micro data.

Technical note 3. Gender Development Index

The Gender Development Index (GDI) measures OECD (2023), UNESCO Institute for Statistics
gender inequalities in achievement in three basic di- (2023) and UNICEF Multiple Indicator Cluster
mensions of human development: health, measured Surveys (various years).
by female and male life expectancy at birth; educa- • Estimated earned income: Human Development
tion, measured by female and male expected years Report Office estimates based on female and male
of schooling for children and female and male mean shares of the economically active population, the
years of schooling for adults ages 25 years and older; ratio of the female to male wage in all sectors and
and command over economic resources, measured gross national income in 2017 purchasing power
by female and male estimated earned income. parity (PPP) terms, and female and male shares of
population from ILO (2023), IMF (2023), UNDESA
Data sources (2023), United Nations Statistics Division (2023)
and World Bank (2023).
• Life expectancy at birth: UNDESA (2022).
• Expected years of schooling: CEDLAS and World Steps to calculate Gender
Bank (2023), ICF Macro Demographic and Health Development Index values
Surveys (various years), UNESCO Institute for
Statistics (2023) and United Nations Children’s There are four steps to calculating GDI values.
Fund (UNICEF) Multiple Indicator Cluster Surveys
(various years). Step 1. Estimating female and male earned incomes
• Mean years of schooling for adults ages 25
and older: Barro and Lee (2018), ICF Macro To calculate estimated earned incomes, the share
Demographic and Health Surveys (various years), of the wage bill is calculated for each gender. The

TECH NICAL NOTES 1–6 7

 female share of the wage bill (Sf) is calculated as Having defined the minimum and maximum val-
follows: ues, the subindices are calculated as follows:

Wf /Wm . EAf actual value – minimum value .

Dimension index =
Sf = maximum value – minimum value
Wf /Wm . EAf + EAm
For education the dimension index is first obtained
where Wf/Wm is the ratio of female to male wage, EAf for each of the two subcomponents, and then the un-
is the female share of the economically active popula- weighted arithmetic mean of the two resulting indi-
tion and EAm is the male share. ces is taken.
The male share of the wage bill is calculated as:
Step 3. Calculating the female and male
Sm = 1 – Sf.
Human Development Index values
Estimated female earned income per capita
(GNIpcf ) is obtained from GNI per capita (GNIpc), The female and male HDI values are the geometric
first by multiplying it by the female share of the wage means of the three dimensional indices for each gender:
bill, Sf , and then rescaling it by the female share of
HDIf = (IHealth . IEducation . IIncome )1/3
the population, Pf = Nf /N: f f f

HDIm = (IHealth . IEducation . IIncome )1/3

GNIpcf = GNIpc . Sf /Pf . m m m

Estimated male earned income per capita is ob-

tained in the same way: Step 4. Comparing female and male
GNIpc = GNIpc . S /P
m m m
Human Development Index values

where Pm = 1 – Pf is the male share of population. The GDI is simply the ratio of female HDI value to
male HDI value:
Step 2. Normalizing the indicators HDIf
GDI = .
HDIm
To construct the female and male HDI values, first
the indicators, which are in different units, are trans- Example: Bangladesh (2022 GDI)
formed into indices and then dimension indices for
each sex are aggregated by taking the geometric mean. Female Male
Indicator value value
The indicators are transformed into indices on a
Life expectancy at birth (years) 76.012 71.514
scale of 0 to 1 using the same goalposts that are used
Expected years of schooling (years) 12.407 11.502
for the HDI, except life expectancy at birth, which is
Mean years of schooling (years) 6.755 8.021
adjusted for the average five-year biological advan-
Wage ratio (female/male) 0.800
tage that women have over men.
Gross national income per capita (2017 PPP $) 6,511.122

Share of economically active population 0.333 0.667

Goalposts for the Gender Development
Share of population 0.504 0.496
Index in this Report

Indicator Minimum Maximum Female wage bill:

Life expectancy at birth (years) Sf = (0.800 ∙ 0.333) / [(0.800 ∙ 0.333) + 0.667] = 0.285
Female 22.5 87.5

Male 17.5 82.5 Estimated female earned income per capita:

Expected years of schooling (years) 0 18 GNIpcf = 6,511.122 ∙ 0.285 / 0.504 = 3,681.884
Mean years of schooling (years) 0 15

Estimated earned income (2017 PPP $) 100 75,000 Male wage bill:
Sm = 1 – 0.285 = 0.715
Note: For the rationale on choice of minimum and maximum values, see Technical
note 1.

8 HUMA N DEVELOP MENT R EP ORT 2023/2024

Estimated male earned income per capita: Gender Development Index groups
GNIpcm = 6,511.122 ∙ 0.715 / 0.496 = 9,385.992
The GDI groups are based on the absolute deviation
Female health index = (76.012 – 22.5) / (87.5 – 22.5) = 0.823 of GDI from gender parity, 100 ∙ |GDI – 1|. Coun-
tries with absolute deviation from gender parity of
Male health index = (71.514 – 17.5) / (82.5 – 17.5) = 0.831 2.5 percent or less are considered countries with high
equality in HDI achievements between women and
Female education index = [(12.407 / 18) + (6.755 / 15)] / 2 = 0.570 men and are classified as group 1. Countries with ab-
solute deviation from gender parity of 2.5–5 percent
Male education index = [(11.502 / 18) + (8.021 / 15)] / 2 = 0.587 are considered countries with medium-high equal-
ity in HDI achievements between women and men
Estimated female earned income index: and are classified as group 2. Countries with abso-
[ln(3,681.884) – ln(100)] / [ln(75,000) – ln(100)] = 0.545 lute deviation from gender parity of 5–7.5 percent are
considered countries with medium equality in HDI
Estimated male earned income index: achievements between women and men and are clas-
[ln(9,385.992) – ln(100)] / [(ln(75,000) – ln(100)] = 0.686 sified as group 3. Countries with absolute deviation
from gender parity of 7.5–10 percent are considered
Female HDI = (0.823 ∙ 0.570 ∙ 0.545)1/3 = 0.6347 countries with medium-low equality in HDI achieve-
ments between women and men and are classified
Male HDI = (0.831 ∙ 0.587 ∙ 0.686)1/3 = 0.6943 as group 4. Countries with absolute deviation from
gender parity of more than 10 percent are considered
GDI = 0.6347 / 0.6943 = 0.914 countries with low equality in HDI achievements be-
Note: Values are rounded. tween women and men and are classified as group 5.

Technical note 4. Gender Inequality Index

The Gender Inequality Index (GII) reflects Data sources
gender-based disadvantage in three dimensions —
reproductive health, empowerment and the labour • Maternal mortality ratio (MMR): WHO, UNICEF,
market—for as many countries as data of reasonable UNFPA, World Bank Group and UNDESA/
quality allow. It shows the loss in potential human Population Division (2023).
development due to inequality between female and • Adolescent birth rate (ABR): UNDESA (2022).
male achievements in these dimensions. It rang- • Share of parliamentary seats held by each sex (PR):
es from 0, where women and men fare equally, to 1, IPU (2023).
where one gender fares as poorly as possible in all • Population with at least some secondary edu-
measured dimensions. cation (SE): Barro and Lee (2018), ICF Macro
GII values are computed using the association-sen- Demographic and Health Surveys (various years),
sitive inequality measure suggested by Seth (2009), OECD (2023), UNESCO Institute for Statistics
which implies that the index is based on the general (2023) and United Nations Children’s Fund
mean of general means of different orders—the first Multiple Indicator Cluster Surveys (various years).
aggregation is by a geometric mean across dimen- • Labour force participation rate (LFPR): ILO (2023).
sions; these means, calculated separately for women
and men, are then aggregated using a harmonic Steps to calculate Gender
mean across genders. Inequality Index values

There are five steps to calculating GII values.

TECH NICAL NOTES 1–6 9

 Step 1. Treating zeros and extreme values Step 4. Calculating the geometric mean of
the arithmetic means for each indicator
Because a geometric mean cannot be computed from
zero values, a minimum value of 0.1 percent is set for The reference standard for computing inequali-
all component indicators. Further, as higher maternal ty is obtained by aggregating female and male indi-
mortality suggests poorer maternal health, for the ma- ces using equal weights (thus treating the genders
ternal mortality ratio the maximum value is truncated equally) and then aggregating the indices across
at 1,000 deaths per 100,000 births and the minimum dimensions:
value at 10. The rationale is that countries where mater-
GF, M = Health . Empowerment . LFPR
3
nal mortality ratios exceed 1,000 do not differ in their
inability to create conditions and support for maternal
health and that countries with 10 or fewer deaths per 10 1
where Health = . + 1 /2,
100,000 births are performing at essentially the same MMR ABR
level and that small differences are random. Sensitivity
analysis of the GII is given in Gaye and others (2010). Empowerment = ( PRF . SEF + PRM . SEM /2 and)
Step 2. Aggregating across dimensions within LFPRF + LFPRM
LFPR = 2 .
each gender group, using geometric means
Health should not be interpreted as an average of
Aggregating across dimensions for each gender corresponding female and male indices but rather
group by the geometric mean makes the GII associa- as half the distance from the norms established for
tion sensitive (see Seth 2009). the reproductive health indicators—fewer maternal
For women and girls the aggregation formula is: deaths and fewer adolescent pregnancies.

GF = 3 10 . 1 1/2
. (PR . SE )1/2 . LFPR ,  (1) Step 5. Comparing indices
MMR ABR F F F

Comparing the equally distributed gender index to

and for men and boys the formula is the reference standard yields the GII,
GM = 1 . (PRM . SEM) 1/2 . LFPRM .
3

HARM (GF , GM )
1– .
The rescaling by 0.1 of the maternal mortality ratio GF,– M–
in equation 1 is needed to account for the truncation
of the maternal mortality ratio at 10. Example: Costa Rica (2022 GII)

Step 3. Aggregating across gender Labour

Health Empowerment market
groups, using a harmonic mean
Adolescent
Maternal birth rate Population
The female and male indices are aggregated by the mortality (births with at
ratio per 1,000 Share of least some
harmonic mean to create the equally distributed gen- (deaths per women seats in secondary Labour force
100,000 ages parliament education participation
der index
live births) 15–19) (% held) (%) rate (%)

(GF)–1 + (GM)–1 –1 Female 22 35.7 47.4 50.2 50.1

HARM (GF , GM) = 2 . Male na na 52.6 48.1 72.9

Using the harmonic mean of within-group ge- F+M

2
( )( )
10
22
1
35.7
+1
= 0.5564
0.474 . 0.502 + 0.526 . 0.481 0.501 + 0.729
2 2
ometric means captures the inequality between 2 = 0.4954 = 0.6150

women and men and adjusts for association between na is not applicable.
dimensions—that is, it accounts for the overlapping
inequalities in dimensions.

10 HUMA N DEVELOP MENT R EP ORT 2023/2024

 Using the above formulas, it is straightforward to obtain:

10 . 1 .
GF : 3 0.474 . 0.502 . 0.501 = 0.3021
22 35.7

GM : 3
1. 0.526 . 0.481 . 0.729 = 0.7158

1 1 1 –1
HARM (GF , GM ): 2 0.3021 + 0.7158 = 0.4249

0.5564 . 0.4954 . 0.6150 = 0.5534

3
GF,– M– :

GII: 1 – (0.4249 / 0.5534) = 0.232

Technical note 5. Multidimensional Poverty Index

The global Multidimensional Poverty Index (MPI) de Salud y Nutrición; Jamaica, 2018 Jamaica Survey
identifies multiple deprivations at the household level of Living Conditions; Libya, 2014 Pan Arab Pop-
in health, education and standard of living. It uses ulation and Family Health Survey; Mexico, 2021
micro data from household surveys, and—unlike the Encuesta Nacional de Salud y Nutricion; Morocco,
Inequality-adjusted Human Development Index— 2017/2018 Pan Arab Population and Family Health
all the indicators needed to construct the measure Survey; Mozambique, 2019/2020 Inquerito Sobre
must come from the same survey. More details about Orcamento Familiar; Peru, 2021 Encuesta De-
the general methodology can be found in Alkire and mográfica y de Salud Familiar; Seychelles, 2019
Jahan (2018). Programmes (Stata do-files) for com- Quarterly Labor Force Survey; and Sri Lanka, 2016
puting the MPI and its components for a large selec- Demographic and Health Survey.
tion of countries with appropriate data are available
at https://hdr.undp.org/mpi-statistical-programmes. Methodology
Programmes for computing the MPI in R format are
also available for a small selection of countries.2 The 2023 global MPI has the same functional form
and indicators as in 2018. It continues to use 10 in-
Data sources dicators in three dimensions—health, education and
standard of living—following the same dimensions
• ICF Macro Demographic and Health Surveys. and weights as the Human Development Index.
• United Nations Children’s Fund Multiple Indicator Each person is assigned a deprivation score accord-
Cluster Surveys. ing to his or her household’s deprivations in each of
• For several countries, national household surveys the 10 indicators. The maximum deprivation score
with the same or similar content and questionnaires: is 100 percent, with each dimension equally weight-
Plurinational State of Bolivia, 2016 Encuesta de De- ed; thus, the maximum deprivation score in each di-
mografía y Salud; Botswana, 2015/2016 Multi-Topic mension is 33.3 percent or, more accurately, ⅓. The
Household Survey; Brazil, 2015 Pesquisa Nacional health and education dimensions have two indicators
por Amostra de Domicílios; China, 2014 China Fam- each, so each indicator is weighted as 1/6. The stand-
ily Panel Studies; Ecuador, 2018 Encuesta Nacional ard of living dimension has six indicators, so each in-
2 These global estimates are jointly produced by the Oxford Poverty and Hu- dicator is weighted as 1/18.
man Development Initiative and the Human Development Report Office using To identify multidimensionally poor people, the
110 survey datasets. For harmonized changes over time, produced by Alkire,
Kanagaratnam and Suppa (2023) using 205 survey datasets, see table 2 in deprivation scores for each indicator are summed to
OPHI and UNDP (2023).

TECH NICAL NOTES 1–6 11

 obtain the household deprivation score. A cutoff of The intensity of poverty, A, reflects the average
⅓ is used to distinguish between poor and nonpoor proportion of the weighted component indicators in
people. If the deprivation score is ⅓ or higher, that which multidimensionally poor people are deprived.
household (and everyone in it) is considered multi- For multidimensionally poor people only (those with
dimensionally poor. People with a deprivation score a deprivation score s greater than or equal to 33.3 per-
of 1/5 or higher but less than ⅓ are considered to be cent), the deprivation scores are summed and divid-
vulnerable to multidimensional poverty. People with ed by the total number of multidimensionally poor
a deprivation score of ½ or higher are considered to people:
be in severe multidimensional poverty. q
∑1 si
The headcount ratio, H, is the proportion of multi- A=
q
dimensionally poor people in the population:
The headcount ratio, H, or incidence of multidi- where si is the deprivation score that the ith multidi-
mensional poverty is the proportion of multidimen- mensionally poor person experiences.
sionally poor people in the population: The deprivation score si of the ith multidimen-
q sionally poor person can be expressed as the
H= n
sum of the weights associated with each indica-
where q is the number of people who are multidimen- tor j ( j = 1, 2, ..., 10) in which person i is deprived,
sionally poor and n is the total population. si = ci1 + ci 2 + … + ci10.

Dimension Indicator Deprived if... Weight

Health Nutrition Any adult under age 70 years or any child for whom nutritional information is available is undernourished. Adults ages 1/6
19–70 years (229–840 months) are considered undernourished if their body mass index (BMI) is below 18.5 kg/m2.
Individuals ages 5–19 years (61–228 months) are considered undernourished if their age-specific BMI values are below
minus two standard deviations from the median of the reference population (https://www.who.int/growthref/en/). In the
majority of countries, BMI-for-age covered individuals ages 15–19 years, as anthropometric data were available only for this
age group; if other data were available, BMI-for-age was applied for all individuals ages 5–19 years. Children under age
5 years (under 60 months) are considered undernourished if their z-score for either height-for-age (stunting) or weight-
for-age (underweight) is below minus two standard deviations from the median of the reference population (https://www​
.who.int/childgrowth/software/en/). Nutritional information is not provided for households without members eligible for
measurement; these households are assumed to be nondeprived in this indicator.
Child mortality Any child under age 18 has died in the five years preceding the survey. When a survey lacks information about the date of 1/6
child deaths, deaths that occurred at any time are taken into account.a
Education Years of No household member of “school entrance age + six years” or older has completed at least six years of schooling.b 1/6
schooling
School Any school-age childc is not attending school up to the age at which he or she would complete class 8. 1/6
attendance
Standard of Electricity The household has no electricity.d 1/18
living Sanitation The household does not have access to improved sanitation (according to Sustainable Development Goal guidelines), or 1/18
it is improved but shared with other households. A household is considered to have access to improved sanitation if it has
some type of flush toilet or latrine or ventilated improved pit or composting toilet that is not shared. When a survey uses a
different definition of adequate sanitation, the survey report is followed.
Drinking water The household does not have access to an improved source of drinking water (according to Sustainable Development 1/18
Goal guidelines), or an improved source of drinking water is at least a 30-minute walk from home, roundtrip. A household
is considered to have access to an improved source of drinking water if the source is piped water, a public tap, a borehole
or pump, a protected well, a protected spring or rainwater. When a survey uses a different definition of improved source
of drinking water, the survey report is followed.
Housing At least one of the household’s three dwelling elements—floor, walls or roof—is made of inadequate materials—that is, the 1/18
floor is made of natural materials and/or the walls and/or the roof are made of natural or rudimentary materials. The floor
is made of natural materials such as mud, clay, earth, sand or dung; the dwelling has no roof or walls; the roof or walls are
constructed using natural materials such as cane, palm, trunks, sod, mud, dirt, grass, reeds, thatch, bamboo or sticks or
rudimentary materials such as carton, plastic or polythene sheeting, bamboo or stone with mud, loosely packed stones,
uncovered adobe, raw or reused wood, plywood, cardboard, unburnt brick, or canvas or tent.
Cooking fuel The household cooks with dung, wood, charcoal or coal. 1/18
Assets The household does not own a car or truck and does not own more than one of the following assets: radio, television, 1/18
telephone, computer, animal cart, bicycle, motorbike or refrigerator.e

a. Information about child deaths is typically reported by women ages 15–49. When information from an eligible woman was not available, information from a man was used
when the man reported no death in the household, and information was coded as missing when the man reported a death (because the date of the death was unknown).
b. The cutoff was previously age 10, but this did not account for the fact that children do not normally complete six years of schooling by age 10. If a child starts school at
age 6 and has completed six years of schooling by age 10, this is counted as an achievement. If not, this is not counted as a deprivation.
c. Official school entrance age is from UIS.Stat (http://data.uis.unesco.org).
d. For the few countries that do not collect data on electricity because of 100 percent coverage, all households in the country as identified as nondeprived in electricity.
e. Televisions include smart televisions and black and white televisions, telephones include cell phones, computers include tablets and laptops, and refrigerators include freezers.

12 HUMA N DEVELOP MENT R EP ORT 2023/2024

 The MPI value is the product of two measures: the (the average multidimensionally poor person is de-
incidence of multidimensional poverty and the inten- prived in 56.3 percent of the weighted indicators).
sity of poverty:
MPI = H . A MPI = H . A = 0.8 . 0.563 = 0.450.
The contribution of dimension d to multidimen-
sional poverty can be expressed as Contribution of deprivations in:
q
∑ j∈d ∑ 1 cij Health:
Contribd = n / MPI
16.67 . 5 + 16.67 . (7 + 4)
contrib1 =
4+7+5+4
/ 0.450 = 29.6%
where d is health, education or standard of living.
The MPI can also be expressed as the weighted Education:
sum of the censored headcount rates hj of each in-
16.67 . (7 + 4) + 16.67 . 7
dicator j. The censored headcount rate of indicator contrib2 =
4+7+5+4
/ 0.450 = 33.3%
j refers to the proportion of people who are multidi-
mensionally poor and deprived in this indicator. Standard of living:

MPI = ∑ j=1 cj . hj contrib3 = 5.56 . (7 . 4 + 5 . 4 + 4 . 3) / 0.450 = 37.1%.

10

where cj is the weight associated with indicator j (ei-

ther 1/6 or 1/18), and the weights sum to 1.
The variance of deprivation scores of multidimen-
sionally poor people is used to measure inequality Example using hypothetical data
among those people:
Indicator Household
q Indicator weight 1 2 3 4
V = ∑ (si – A)2 / (q – 1) Household size 4 7 5 4
1
Health

All parameters defined above are estimated using At least one member is
undernourished ( 1/3) ÷ 2 = 16.7% 0 0 1 0
survey data and sampling weights according to the One or more children have died (1/3) ÷ 2 = 16.7% 1 1 0 1
rules of the sampling theory (Lohr 2010). Education
Weighted deprivations: No one has completed at least six
years of schooling (1/3) ÷ 2 = 16.7% 0 1 0 1
• Household 1: (1 . 16.67) + (1 . 5.56) = 22.2 percent.
At least one school-age child not
• Household 2: 72.2 percent. enrolled in school (1/3) ÷ 2 = 16.7% 0 1 0 0

• Household 3: 38.9 percent. Standard of living

No electricity (1/3) ÷ 6 = 5.6% 0 1 1 1
• Household 4: 50.0 percent.
No access to improved sanitation (1/3) ÷ 6 = 5.6% 0 0 1 0
Based on this hypothetical population of four No access to an improved source of
households: drinking water (1/3) ÷ 6 = 5.6% 0 1 1 0
House built with inadequate materials (1/3) ÷ 6 = 5.6% 0 0 0 0
Household cooks with dung, wood,
Incidence of multidimensional poverty (H) = charcoal or coal (1/3) ÷ 6 = 5.6% 1 1 1 1
Household does not own a car or truck
0+7+5+4 and does not own more than one of
4 + 7 + 5 + 4 = 0.80 the following assets: radio, television,
telephone, computer, animal cart,
bicycle, motorbike or refrigerator. (1/3) ÷ 6 = 5.6% 0 1 0 1
(80 percent of people are multidimensionally poor).
Results
Individual deprivation score, c (sum
Intensity of poverty (A) = of each deprivation multiplied by its
weight) for each household member 22.2% 72.2% 38.9% 50.0%

(72.2 . 7) + (38.9 . 5) + (50.0 . 4) Is the household multidimensionally

= 56.3 percent poor (c ≥ 1/3)? No Yes Yes Yes
(7+5+4)
Note: 1 indicates deprivation in the indicator; 0 indicates nondeprivation.

TECH NICAL NOTES 1–6 13

 Calculating the contribution of each dimension to Censored headcount rates for each indicator
multidimensional poverty provides information that
can be useful for revealing a country’s deprivation Proportion of
people who are
structure and can help with policy targeting. Proportion of ­multidimensionally
People who are people who are poor and deprived
multidimensionally ­multidimensionally in each indicator
Variance of deprivation scores among the poor poor and deprived poor and deprived multiplied by the
(V) = in each indicator in each indicator indicator weight
Nutrition 5 0.250 0.042
(0.722 – 0.563)2 ∙ 7 + (0.389 – 0.563)2 ∙ 5 + (0.500 – 0.563)2 ∙ 4 Child mortality 11 0.550 0.092
16 – 1 Years of
= 0.023 schooling 11 0.550 0.092
School
attendance 7 0.350 0.058
Electricity 16 0.800 0.044
Sanitation 5 0.250 0.014
Drinking water 12 0.600 0.033
Housing 0 0.000 0.000
Cooking fuel 16 0.800 0.044
Assets 11 0.550 0.031
Sum 0.450

Technical note 6. Planetary pressures–adjusted Human

Development Index

The Planetary pressures–adjusted Human Devel- Indicator definitions and data sources
opment Index (PHDI) adjusts the Human Devel-
opment Index (HDI) for planetary pressures in the In addition to the data used to calculate the HDI,
Anthropocene. The PHDI discounts the HDI for the PHDI uses data on carbon dioxide emissions per
pressures on the planet to reflect a concern for in- capita (production) and material footprint per capita.
tergenerational inequality, similar to the Inequal- • Carbon dioxide emissions per capita (production):
ity-adjusted HDI adjustment, which is motivated carbon dioxide emissions produced as a conse-
by a concern for intragenerational inequality. The quence of human activities (use of coal, oil and gas
PHDI is computed as the product of the HDI and (1 for combustion and industrial processes, gas flar-
– index of planetary pressures), where (1 – index of ing and cement manufacture), divided by midyear
planetary pressures) can be seen as an adjustment population. Values are territorial emissions, mean-
factor. ing that emissions are attributed to the country in
The PHDI is the level of human development which they physically occur. Data are from Global
adjusted by carbon dioxide emissions per capita Carbon Project (2023).
(production) and material footprint per capita to • Material footprint per capita: material footprint
account for the excessive human pressure on the is the attribution of global material extraction to
planet. It should be seen as an incentive for trans- domestic final demand of a country. Material foot-
formation. In an ideal scenario, with no pressures print is calculated as raw material equivalent of im-
on the planet, the PHDI equals the HDI. Howev- ports plus domestic extraction minus raw material
er, as pressures increase, the PHDI falls below the equivalents of exports. The total material footprint
HDI. In this sense the PHDI measures the level of is the sum of the material footprint for biomass,
human development when planetary pressures are fossil fuels, metal ores and nonmetal ores. Material
considered. footprint per capita describes the average material
use for final demand. Data are from UNEP (2023).

14 HUMA N DEVELOP MENT R EP ORT 2023/2024

Steps to calculate Planetary Step 3. Adjusting the Human Development
pressures-adjusted Human Index to account for planetary pressures
Development Index values
The PHDI is the product of the HDI and the adjust-
There are three steps to calculating PHDI values. ment factor, A:

Step 1. Calculating the carbon dioxide emissions PHDI = HDI . A,

index and the material footprint index
or, equivalently, PHDI = HDI . (1 – P).
Carbon dioxide emissions per capita and material The difference between the HDI and the PHDI val-
footprint per capita are normalized in the same way ues due to planetary pressures, expressed as a per-
as the components of the HDI. Through a min-max centage, is:
transformation each becomes an index with values
HDI – PHDI
between 0 and 1 calculated as: Difference in HDI value = . 100
HDI
Aj index = (maximumj – observed valuej) /
= P . 100
(maximumj – minimumj)

where j = 1,2 refers to the two included planetary

pressure indicators. Example: Azerbaijan (2022 PHDI)
Zero was set as minimum. The maximum corre-
sponds to the maximum value observed historically Indicator Value

for all countries since 1990, in line with the similar ap- Human Development Index (HDI) 0.760

proaches in the literature, such as Biggeri and Mauro Carbon dioxide emissions per capita 3.7
(production), tonnes
(2018). For carbon dioxide emissions per capita the
Material footprint per capita, tonnes 8.3
maximum value is 76.61 tonnes, observed for Qatar in
Carbon dioxide emissions index (76.61 – 3.7) / 76.61 =
1997. For material footprint per capita the maximum 0.951
value is 140.82, observed for Qatar in 2008. The ranking Material footprint index (140.82 – 8.3) / 140.82
of countries is sensitive to the selection of the maximum. = 0.941

For both carbon dioxide emissions per capita and Adjustment for planetary pressures (0.951 + 0.941) / 2 =
factor (A) 0.946
material footprint per capita, the higher the observed
Planetary pressures–adjusted HDI (PHDI) 0.760 ∙ 0.946 = 0.719
value and the closer to the maximum, the higher the
Difference in HDI value (%) [(0.760 – 0.719) / 0.760] ∙
pressures on the planet, implying a smaller value of 100 = 5.4
the index and a larger adjustment to the HDI.
Note: Values are rounded.

Step 2. Constructing the adjustment

for planetary pressures Difference in Planetary
pressures–adjusted Human
The adjustment factor for planetary pressures (A) is Development Index rank and
the arithmetic average of the indices measuring car- Human Development Index rank
bon dioxide emissions per capita and material foot-
print per capita, which assumes perfect substitution Difference in ranks on the PHDI and the HDI is cal-
of these two indicators. Lower pressures on the planet culated only for countries for which both a PHDI and
result in a larger A and smaller adjustment to the HDI. an HDI value are calculated. Due to data limitations,
PHDI values are calculated for a smaller set of coun-
A = (Carbon dioxide emissions index +
tries. To calculate the PHDI rank difference from the
material footprint index) / 2
HDI rank, HDI ranks are recalculated for the set of
In addition, the index of planetary pressures, P, is countries for which a PHDI value is calculated.
defined as the complement of A: P = (1 – A).

TECH NICAL NOTES 1–6 15

References
Alkire, S., and S. Jahan. 2018. “The New Global MPI 2018: Aligning with the Lohr, S. 2010. Sampling: Design and Analysis, 2nd ed. Boston, MA: Brooks/Cole.
Sustainable Development Goals.” Human Development Research Paper. UNDP-
HDRO, New York. http://hdr.undp.org/en/content/new-global-mpi-2018-aligning​ Maddison, A. 2010. Historical Statistics of the World Economy, 1–2030 AD. Paris:
-sustainable-development-goals. Organisation for Economic Co-operation and Development.

Anand, S., and A. Sen. 2000. “The Income Component of the Human Development OECD (Organisation for Economic Co‑operation and Development). 2023. OECD.
Index.” Journal of Human Development and Capabilities (1)1: 83–106. Stat. https://stats.oecd.org. Accessed 20 September 2023.

Alkire, S., Kanagaratnam, U., and Suppa, N. 2023. “A Methodological Note on Oeppen, J., and J.W. Vaupel. 2002. “Broken Limits to Life Expectancy.” Science
the Global Multidimensional Poverty Index (MPI) 2023 Changes over Time Results 296: 1029–1031.
for 84 Countries.” OPHI MPI Methodological Note 57, Oxford Poverty and Human
Development Initiative, University of Oxford, UK. OPHI (Oxford Poverty and Human Development Initiative) and UNDP (United
Nations Development Programme). 2023. Global Multidimensional Poverty Index
Atkinson, A. 1970. “On the Measurement of Economic Inequality.” Journal of Eco- 2023: Unstacking Global Poverty: Data for High Impact Action. New York and Ox-
nomic Theory 2(3): 244–263. ford, UK.

Barro, R.J., and J.-W. Lee. 2018. Dataset of Educational Attainment, June 2018 Revi- Riley, J.C. 2005. Poverty and Life Expectancy. Cambridge, UK: Cambridge Univer-
sion. www.barrolee.com. Accessed 9 August 2023. sity Press.

Biggeri, M., and Mauro, V. 2018. “Towards a More ‘Sustainable’ Human Development Seth, S. 2009. “Inequality, Interactions, and Human Development.” Journal of Hu-
Index: Integrating the Environment and Freedom.” Ecological Indicators 91: 220–231. man Development and Capabilities 10(3): 375–396.

CEDLAS (Center for Distributive, Labor and Social Studies) and World Bank. UNDESA (United Nations Department of Economic and Social Affairs). 2022.
2023. Socio-Economic Database for Latin America and the Caribbean. http://www​ World Population Prospects: The 2022 Revision. New York. https://population.un​
.cedlas.econo.unlp.edu.ar/wp/en/estadisticas/sedlac/estadisticas. Accessed 19 .org/wpp/. Accessed 1 August 2023.
September 2023.
UNDESA (United Nations Department of Economic and Social Affairs). 2023.
Eurostat. 2023. European Union Statistics on Income and Living Conditions. EU- World Economic Situation and Prospects 2023. New York. https://www.un.org​
SILC UDB 2021 – version of October 2023. Brussels. https://ec.europa.eu/eurostat​ /development/desa/dpad/publication/world-economic-situation-and-prospects​
/web/microdata/european-union-statistics-on-income-and-living-conditions. Ac- -2023/. Accessed 15 November 2023.
cessed 30 October 2023.
UNEP (United Nations Environment Programme). 2023. International Resource
Foster, J., L. Lopez-Calva and M. Szekely. 2005. “Measuring the Distribution of Hu- Panel’s Global Material Flows Database. https://www.resourcepanel.org/global​
man Development: Methodology and an Application in Mexico.” Journal of Human -material-flows-database. Accessed 18 December 2023.
Development and Capabilities 6(1): 5–25.
UNESCO (United Nations Educational, Scientific and Cultural Organization) In-
Gaye, A., J. Klugman, M. Kovacevic, S. Twigg and E. Zambrano. 2010. “Measur- stitute for Statistics. 2023. UIS Developer Portal, Bulk Data Download Service.
ing Key Disparities in Human Development: The Gender Inequality Index.” Human https://apiportal.uis.unesco.org/bdds. Accessed 19 September 2023.
Development Research Paper. UNDP-HDRO, New York http://hdr.undp.org/sites​
/default/files/hdrp_2010_46.pdf. UNICEF (United Nations Children’s Fund). Various years. Multiple Indicator Clus-
ter Surveys. New York. http://mics.unicef.org/. Accessed 31 October 2023.
Global Carbon Project. 2023. Global Carbon Atlas. https://globalcarbonatlas.org​
/emissions/carbon-emissions/. Accessed 24 November 2023. United Nations Statistics Division. 2023. National Accounts Main Aggregates Da-
tabase. http://unstats.un.org/unsd/snaama. Accessed 15 November 2023.
ICF Macro. Various years. D
 emographic and Health Surveys. http://dhsprogram​
.com. Accessed 28 November 2023. UNU-WIDER (United Nations University World Institute for Development Eco-
nomics Research). 2023. World Income Inequality Database Companion dataset.
ILO (International Labour Organization). 2023. ILOSTAT database. https://­ilostat​ Version 28 November 2023. https://www.wider.unu.edu/database/world-income​
.ilo.org/data/. Accessed 22 November 2023. -inequality-database-wiid#WIIDcomp. Accessed 28 November 2023.

IMF (International Monetary Fund). 2023. World Economic Outlook database. WHO (World Health Organization), UNICEF (United Nations Children’s Fund),
Washington, DC. https://www.imf.org/en/Publications/WEO/weo-database/2023​ UNFPA (United Nations Population Fund), World Bank Group and United Na­
/October. Accessed 15 November 2023. tions Population Division. 2023. Trends in Maternal Mortality 2000 to 2020: Esti-
mates by WHO, UNICEF, UNFPA, World Bank Group and UNDESA/Population Di-
IPU (Inter-Parliamentary Union). 2023. Parline database: Monthly ranking of women vision. Geneva: World Health Organization. https://www.who.int/publications/i/item​
in national parliaments. https://data.ipu.org/women-ranking. Accessed 1 August 2023. /9789240068759. Accessed 1 August 2023.

LIS. 2023. Luxembourg Income Study Database. http://www.lisdatacenter.org/data​ World Bank. 2023. World Development Indicators database. Washington, DC.
-access. Accessed 30 September 2023. http://data.worldbank.org. Accessed 7 November 2023.

16 HUMA N DEVELOP MENT R EP ORT 2023/2024